Monday, June 29, 2026 | 11:19 PM ISTहिंदी में पढें
Business Standard
Notification Icon
userprofile IconSearch

Exl sees iMerit as gateway to foundation model training: Rohit Kapoor

EXL says its $310-million iMerit acquisition will strengthen foundation model training capabilities as enterprises increasingly adopt AI-driven business operations

rohit kapoor
premium

AI needs as much capability on the process side as it does on the technology side- Rohit Kapoor, Chairman & CEO, ExlService Holdings

Avik Das

Listen to This Article

ExlService Holdings, which bought artificial intelligence (AI) model-training firm iMerit last week for $310 million, says the latest acquisition will help it offer a wider range of services to clients beyond data management. Business process management (BPM) companies have been pivoting to focus more on data analytics and AI rather than low-value work, which is being automated rapidly. In a video conversation with Avik Das, Chairman and Chief Executive Officer Rohit Kapoor, joining from New York, talks about how BPM companies are growing faster than information technology (IT) companies, the demand environment, and acquisition strategies. Edited excerpts:
 
What is the reason for buying iMerit? 
Exl already has a strong capability in data management. But when it comes to models, this is our first foray into acquiring a capability that will help train foundation models. Our view is that enterprises are going to shift towards adopting specialised language models and small language models.
 
And so we will be able to evaluate these models, train them, and make them much more effective for the enterprise. The biggest issue that enterprises have is that they invest in AI but don’t get the returns from it. That’s because the model doesn’t understand the context. To train these models, you need a methodology and a platform. This acquisition will allow us to evaluate, fine-tune, and improve these models so that they are better suited for enterprise use.
 
How is the demand environment, and what is the message from clients halfway into this year? 
If you look at our portfolio, 60 per cent of it is entirely focused on data, analytics, and AI, and there we are seeing very strong demand. In fact, we are struggling to find the talent needed to fulfil that demand. So the demand side of that equation is very strong, but not the supply side. Our digital operations business is also growing. We are seeing broad-based strength across the board.
 
The recent growth of BPM companies or the sector seems to be the antithesis of IT services companies. Why do you think this has happened? 
IT services companies are getting constrained because there is a lot of compression in the software development lifecycle, which is a large part of their portfolio. For BPM companies, software development is not an issue for us at all. With the adoption of AI, it’s the first time that you have a technology that requires as much capability on the process side and an understanding of workflows and business as it does on the technology side. It’s the businesses that are driving the adoption of AI in the enterprise. So BPM companies are much better positioned on this compared with any other player.
 
If we take a look at all the work that we do around BPM, that’s growing at 10 per cent for us. I think the other thing that is happening is that mid-sized and small-sized companies are able to make the pivot much faster than some of the larger companies. For example, we have been able to make the pivot towards data, AI, and embedding intelligence into operations much ahead of many others. For some of the larger players, that’s probably going to take some time.
 
Can you highlight the demand vectors driving your AI business? 
The very basic thing that every organisation needs is access to its data in order to apply AI. Today, data sits in different silos and platforms within an enterprise. They also have large amounts of contextual and unstructured data that are simply not available for AI use.
 
The second bucket is AI in operations, where we embed AI and agentic AI directly into business processes to improve efficiency and outcomes. At the same time, we redesign workflows so organisations can fully realise the value of AI-enabled operations.
 
The third area is AI services. Many clients do not necessarily want to outsource their operations, but they do want help implementing AI. We support them by building agentic AI platforms, developing AI agents, and embedding AI into their existing workflows.
 
Finally, there is AI solutions, where we have built proprietary industry-specific platforms. For example, in healthcare, our payment integrity solution helps identify fraud opportunities.
 
Will acquisitions continue to be an integral part of your AI growth strategy? 
We will continue to follow a balanced strategy — building organically while pursuing targeted acquisitions where they strengthen our capabilities. Our four AI growth vectors represent areas where we will continue investing organically. From an acquisition perspective, we are interested in AI engineering talent, visual computing capabilities, and expertise in working with multimodal datasets.
 
You have talked about AI reducing certain roles while also creating new job opportunities. How do you see that evolving in the next few years? 
If you look at the past five years, Exl has delivered double-digit revenue growth while growing headcount at 7–9 per cent annually. We continue to expect headcount to increase in the same range.